Comparative analysis of severe infection risk and survival outcomes in DLBCL patients treated with CAR T-cell therapy versus standard chemoimmunotherapy: Insights from real-world data

Abstract Background: Diffuse large B-cell lymphoma (DLBCL) treatments, including standard chemoimmunotherapy (R-CHOP/R-EPOCH) and CAR T-cell therapy, can lead to significant immune dysfunction and infection risk. While CAR-T therapy has improved outcomes for relapsed/refractory DLBCL, its unique toxicities may predispose patients to severe infections. Severe infections such as sepsis or septic shock are a major cause of treatment-related morbidity and mortality in this population. However, there is limited data directly comparing infection outcomes between CAR-T recipients and patients receiving standard chemoimmunotherapy. Objective: To compare the incidence of severe infections and survival outcomes in DLBCL patients treated with CAR T-cell therapy versus those receiving R-CHOP/R-EPOCH chemoimmunotherapy, using contemporary real-world data from 2017 onwards. Methods: We conducted a retrospective cohort study using the TriNetX federated electronic health record network. Adult DLBCL patients (18-80 years) were stratified into two cohorts: CAR T-cell therapy and R-CHOP/R-EPOCH chemoimmunotherapy. Patients with prior malignancies or stem-cell transplants were excluded. Outcomes were assessed over a 12-month follow-up period. The primary outcome was severe infection, defined as sepsis or septic shock. Propensity score matching (1:1) was applied to balance baseline characteristics. Key variables for matching included demographics (age, sex, race/ethnicity) and relevant pre-index comorbidities (e.g., diabetes, chronic kidney disease, baseline neutropenia or cytopenias) to account for differing risk profiles. We calculated the 1-year cumulative incidence of severe infection and death in each matched cohort and used Kaplan–Meier analysis with log-rank tests and Cox proportional hazards models to compare outcomes. Results: The analysis included 1,580 CAR T recipients and 1,580 R-CHOP/R-EPOCH patients after propensity score matching. Baseline demographics and comorbidities were well-balanced between cohorts (median age ~60 years, 36% female, similar rates of diabetes, renal disease, and prior neutropenia in each group). Severe infections (sepsis or septic shock) within one year were slightly more frequent in the R-CHOP/R-EPOCH cohort. The cumulative 1-year incidence of severe infection was approximately 10.5% in CAR T patients vs. 13% in R-CHOP/R-EPOCH patients. This translated to a 2.5% absolute risk reduction in the CAR-T group (risk ratio 0.807, p = 0.042), indicating that standard chemoimmunotherapy patients had a higher risk of developing sepsis during the follow-up of 1 year. Kaplan–Meier analysis showed a comparable survival probability of patients with sepsis in both groups (87% vs. 86.7% at 12 months, Hazard Ratio=0.922, 95% CI 0.737-1.152). Conclusion: In this large real-world analysis of DLBCL patients from 2017 onward, CAR T-cell therapy was not associated with a higher risk of severe infection compared to standard R-CHOP/R-EPOCH chemoimmunotherapy; in fact, the CAR-T cohort experienced a slightly lower 1-year sepsis incidence. Importantly, one-year survival probability in patients with outcome sepsis was comparable between the matched CAR-T and R-CHOP/R-EPOCH groups. These findings highlight that, under contemporary practice conditions, CAR-T therapy's safety profile with respect to life-threatening infections is at least equivalent to (and possibly better than) traditional chemoimmunotherapy. Clinicians can be reassured that adopting CAR-T in eligible DLBCL patients does not incur an excess early infection risk relative to standard therapy. Nonetheless, the substantial infection rates observed in both cohorts underscore the need for vigilant infection prophylaxis and supportive care in all DLBCL patients receiving intensive treatments. Overall, this study provides evidence that CAR-T therapy, despite being used in more advanced disease settings, achieves infection outcomes and survival probabilities comparable to frontline treatment, supporting its safe integration into clinical practice for relapsed/refractory DLBCL patients.